Dr Daniel Susskind, Oxford University economist and author, tells organisations at Advanced World to rethink AI’s impact on the workforce
An Oxford University economist and author has said businesses are trapped in the mindset of AI instantly destroying jobs. Speaking at Advanced World1 last week, Dr Daniel Susskind argues that machines are not trying to replicate the way humans perform certain tasks and that AI will have a more gradual and subtle impact on the workforce.
“There’s this AI fallacy where we believe the only way machines can be intelligent is through replicating human skills,” Daniel explains during his keynote ‘A World Without Work’. “The reality is most jobs can’t be fully automated. According to McKinsey, just 5% of occupations can be 100% automated yet people are focusing on just this headline figure. The same research also shows about 60% of all occupations have at least 30% of tasks that could be automated.
“So, rather than talking about AI replacing jobs, we should actually be breaking these jobs down into tasks and determining which of these can be automated through AI. More often than not, the tasks that can be automated are often routine and do not require creativity.”
During his keynote, Daniel also shared some of the defining moments in AI’s history including the first truly successful form of an AI system, Expert Systems, in the 1980s. And how, following the so-called ‘AI winter’ in the 1990s, AI experienced somewhat of a resurgence when chess player Garry Kasparov was beaten by supercomputer Deep Blue in 1997 and IBM Watson outplayed two humans at the popular TV quiz show Jeopardy! in 2011.
Today, AI is capable of performing tasks for professionals and organisations in a range of industries. It can identify skin cancer more accurately than dermatologists, for example, and it can even design a concert hall – with Hamburg's Elbphilharmonie case in point.
Daniel adds: “We’ve seen AI impact blue-collar workers in industries like agriculture and manufacturing with a phenomenal increase in output and decrease in employment. Now, white-collar workers are equally at risk as there are tasks that can be automated. However, we are not going to see mass unemployment. AI performs tasks fundamentally differently from humans. Ultimately, it has ways of being smart that aren’t smart like us.”
The keynote concluded with the role of lifelong education and what opportunities will exist as adoption of AI increases. A key takeaway was to not train people to do tasks that will be automated but, instead, train them to build, set rules for and monitor AI tools.
Gordon Wilson, CEO at Advanced, concludes: “AI is a challenge but it’s also an opportunity. As Daniel rightly says, we need to look at the tasks, not the jobs. AI can automate tasks to free up people’s time to focus on non-routine tasks – and those that require creativity and judgement – which is especially valuable in industries like the NHS where there is a major shortage of medical professionals. What’s more, by giving time back to the workforce, they are able to better serve their clients – whether patients, students or customers.
“At Advanced, we are working on an AI tool for GPs that will mean most correspondence will no longer require human intervention. As much as 80% of correspondence sent to the GP does not actually need to be read, so AI’s impact on time (and indeed cost) is massive. The tool is currently being piloted at 231 GP practices across the UK and with great results. But this doesn’t mean the role of the GP will become obsolete – far from it.”